Set a measurable task
Choose the hold routine and a requested duration such as 1,000 milliseconds. Keep the pointer within the capture area, start recording and deliberately hold the chosen button before releasing it. Repeat a small number of times with pauses, rather than mixing the test with rapid taps.
The requested duration is part of the report. Completed holds below that value are highlighted, including any that you intentionally released early. The application cannot know your physical intention from an event sequence, so the label remains a review candidate.
Understand interrupted sessions
Leaving the capture area, switching focus, hiding the page, Escape or the session limit stops capture. If a button is still considered held at that point, it is shown as held at the end. The application does not fabricate a release event or use the stop time as a completed hold.
A release without a prior down can occur when capture starts or stops around another interaction. It remains visible as an unpaired release. Review the end reason alongside the timing before interpreting a short or missing hold.
Compare one variable at a time
Keep the same requested hold and rapid-repeat settings when preparing a second session. Use a similar number of trials and the same logical button. A baseline provides a convenient side-by-side summary while the JSON files preserve the detailed sequence.
Browser evidence is downstream of hardware, firmware, input drivers and operating-system behavior. If the observations are inconsistent, retain the files and describe your exact routine. Do not present the tool’s threshold as a manufacturer’s specification.